www.nature.com/scientificreports

OPEN

Body mass index and risk of subtypes of head-neck cancer: the Netherlands Cohort Study

received: 29 May 2015 accepted: 05 November 2015 Published: 04 December 2015

Denise H. E. Maasland1, Piet A. van den Brandt1, Bernd Kremer2 & Leo J. Schouten1 Low body mass index (BMI) has been associated with risk of head-neck cancer (HNC), but prospective data are scarce. We investigated the association between BMI, BMI at age 20 years and change in BMI during adulthood with risk of HNC and HNC subtypes. 120,852 participants completed a questionnaire on diet and other cancer risk factors, including anthropometric measurements, at baseline in 1986. After 20.3 years of follow-up, 411 HNC (127 oral cavity cancer (OCC), 84 oro-/ hypopharyngeal cancer (OHPC), and 197 laryngeal cancer (LC)) cases and 3,980 subcohort members were available for case-cohort analysis using Cox proportional hazards models. BMI at baseline was inversely associated with risk of HNC overall, with a multivariate rate ratio of 3.31 (95% CI 1.40–7.82) for subjects with a BMI  2

378

24.8

2.8

Cases

a

P value

Year of follow-up 0.96c

Categories year of follow-up  0–2

33

24.9

2.2

  > 2–4

40

24.8

3.0

  > 4–6

49

24.8

3.1

  > 6–8

44

24.7

2.4

  > 8–10

47

24.6

2.9

  > 10–12

39

24.6

3.2

  > 12–14

44

24.8

2.9

  > 14–16

47

25.1

2.6

  > 16–18

37

25.2

2.5

  > 18

31

24.9

2.3

Table 2.  BMI in head-neck cancer (HNC) cases according to sex and time between baseline and HNC diagnosis; Netherlands Cohort Study, 1986–2006. aMean ±  standard deviation (SD) of BMI at baseline in subcohort members were 24.9 ±  2.6 kg/m2 for men (N =  1,954) and 25.0 ±  3.5 kg/m2 for women (N =  2,026). b HNC: head-neck cancer; OCC: oral cavity cancer; OHPC: oro-/hypopharyngeal cancer; LC: laryngeal cancer. cT-test of mean BMI at baseline in first two years of follow-up vs. rest of follow-up years.

Cohort Study (NLCS). In addition to BMI at study baseline, we also studied the effects of BMI at age 20 years and change in BMI during adulthood on HNC risk. Finally, we investigated the association of BMI with HNC risk according to smoking status and alcohol consumption.

Results

The mean BMI at baseline of subcohort members (25.0 kg/m2) and cases (24.8 kg/m2) was slightly lower in HNC cases (Table 1). There was a minor difference between subcohort members and cases regarding BMI at age 20 years (21.5 and 21.9 kg/m2, respectively), as well as with respect to change in BMI since age 20 years (plus 3.5 and 3.0 kg/m2, respectively). Among subcohort members and cases, men generally had a similar mean BMI, whereas female cases had a considerably lower BMI at baseline than both male cases and female subcohort members. Notable characteristics with regard to cigarette smoking and alcohol consumption have been described previously12. To examine possible reverse causation, we evaluated BMI at baseline of HNC cases during the follow-up period. As we expected, there was no clear pattern in BMI at baseline among HNC cases diagnosed in the course of 20.3 years of follow-up (Table  2). HNC overall cases diagnosed after the second year of follow-up (N =  378) had a mean BMI at baseline of 24.8 kg/m2, whereas cases diagnosed during the first two years of follow-up (N =  33) had a mean BMI of 24.9 kg/m2, a non-statistically significant difference. Results from age- and sex-adjusted and multivariable-adjusted analyses showed mostly inverse associations between BMI at baseline and risk of HNC overall and HNC subtypes, although these associations were generally somewhat weaker in multivariable-adjusted analyses (Table 3). BMI at baseline was inversely associated with risk of HNC overall, with a multivariate rate ratio (RR) of 3.31 (95% confidence interval (CI) 1.40–7.82) for subjects with a BMI